arX

iv:1

403.

3633

v1 [

phys

ics.

acc-

ph]

14

Mar

201

4

Generating 10∼40MeV high quality monoenergetic electron beams using a 5TW 60fs

laser at Tsinghua University∗

J. F. Hua,† L. X. Yan, C. H. Pai, C. J. Zhang, F. Li, Y. Wan, Y. P. Wu,

X. L. Xu, Y. C. Du, W. H. Huang, H. B. Chen, C. X. Tang, and W. LuDepartment of Engineering Physics,

Tsinghua University, Beijing 100084, China

(Dated: February 14, 2014)

A unique facility for laser plasma physics and advanced accelerator research has been built re-cently at Tsinghua Universtiy. This system is based on Tsinghua Thomson scattering X-ray source(TTX)[1, 2], which combining an ultrafast TW laser with a synchronized 45MeV high brightnesslinac. In our recent laser wakefield acceleration experiments, we have obtained 10∼40MeV highquality monoenergetic electron beams by running the laser at 5TW peak power. Under certainconditions, very low relative energy spread of a few percent can be achieved. Absolute charge cal-ibration for three different scintillating screens has also been performed using the linac system.Anarticle usually includes an abstract, a concise summary of the work covered at length in the mainbody of the article.

keyword: laser-plasma accelerator, monoenergetic, electron charge

PACS numbers: 52.38.Kd, 52.59.Bi, 41.75.Jv

I. INTRODUCTION

Plasma acceleration based on wakefield concept[3, 4]has made great strides in the past decade. By utilizingtable-top 10TW-PW laser systems, high quality monoen-ergetic electron beams with energies up to few GeV[5–11] and energy spread of a few percent[12–14] have beenobtained worldwide. In this paper, we report our re-cent experimental results on laser plasma accelerationutilizing a newly built facility at Tsinghua University.This facility is based on Tsinghua Thomson scatteringX-ray source (TTX), which combining an ultra-fast TWTi:sapphire laser system with a synchronized 45MeV highbrightness linac. By running the laser at 5TW 60fs, highquality 10∼40MeV monoenergetic electron beams havebeen obtained. Under certain plasma conditions, verylow relative energy spread of a few percent has also beenachieved. The paper is organized as follows: in sectionII, the facility will be briefly overviewed; in section III,the experimental results will be presented in details: asummary is provided in section IV.

II. OVERVIEW OF TSINGHUA LASER

PLASMA ACCELERATION PLATFORM

The laser plasma acceleration platform at TsinghuaUniversity has three major components: the TW lasersystem, the 45MeV high brightness linac, and an exper-imental system combining a high vacuum chamber with

∗ Supported by National Natural Science Foundation of China (No.11005063, No. 11375006 and No. 11175102) and Beijing HigherEducation Young Elite Teacher Project

† jfhua@tsinghua.edu.cn

a set of various optical and electric diagnostics. We nextbriefly describe these subsystems in details.

A. TW laser system

The TW laser system (Fig. 1), based on the standardCPA (Chirped Pulse Amplification) and MOPA (Mas-ter Oscillator Power Amplifier) architecture, has six ma-jor components: an oscillator, a stretcher, a regenerativeamplifier, two multi-pass amplifiers and a vacuum com-pressor [15].The 20fs Ti:sapphire oscillator with a rep-rate of

79.3MHz is pumped by a frequency-doubled YVO4 laser(Verdi-5). A fused-silica prism pair is used for intracav-ity dispersion compensation. The synchronization be-tween the oscillator and the RF signal is achieved by apiezoelectric transducer (PZT) based feedback loop withtiming jitter less than 200fs. Before entering the regen-erative amplifier, the seed pulse goes through a Pock-els cell pulse selector and then is stretched to about300ps with a modified offener type stretcher. In the re-generative amplifier pumped by a frequency-doubled Q-switched Nd:YLF laser (Evolution-30), the laser pulseis amplified to about 3mJ with a rep-rate of 1kHz, andthe bandwidth is reduced to 25nm. Another Pockels cellpulse cleaner is inserted before the first multi-pass ampli-fier to enhance the laser contrast against the ns prepulseand ASE. In the first four-pass amplifier pumped by afrequency-doubled Q-switched Nd: YAG lasers (Quanta-Ray Pro-350), the pulse is amplified to 100mJ in a bow-tie configuration. The laser beam is then split into twobeams, with one beam (70mJ) as a seed for the final am-plifier and another beam (30mJ) for third-order harmonicgeneration to drive the RF photocathode gun. The finalamplifier is another four-pass bow-tie amplifier pumped

2

FIG. 1. The ultrashort TW laser system.

with two frequency-doubled Q-switched Nd:YAG lasers(Saga 230) from both ends. The amplification media isa 20mm-long, 20mm-diameter, normal-cut Ti:sapphirecrystal with anti-reflective coating. With a combinedpump energy of 2.8J, the laser pulse can be amplifiedto 1.0 J. A computer-controlled energy tuner made of ahalf-wave plate and a thin-film polarizer is utilized to ad-just the laser pulse energy before compression. The laserpulse is expanded to 50mm-diameter beam size, and thenis sent into the vacuum compressor. After compression,about 70% of the laser energy can be sent into the exper-imental system. The final energy on target can reach upto 500mJ after taking into account the losses on trans-mission.

For laser-plasma interaction experiments, laser tem-poral contrast, focal spot quality and pulse durationare three crucial parameters. High temporal contrast isneeded to avoid pre-plasma ionized by prepulses. Highquality focal spot helps to excite the stable wake and toincrease the energy coupling. Pulse duration should becompressed to as short as possible to increase the peaklaser power. With the significant efforts in the past fewyears, our laser system has been thoroughly optimized forwakefield experiments. Typical parameters of our laserare listed in Tab. I.

The contrast at the nanosecond is measured by a fastphotodiode coupled with ND filters, and the contrastat the picosecond is measured by a commercial third-order autocorrelator. The contrast has also been verifiedthrough experiments by observing pre-plasma using in-terferometry. At laser peak power, no pre-plasma hasbeen found by scanning the delay time between the mainlaser pulse and the probe pulse. To further enhancethe contrast, a technology called cross-polarized wave(XPW) generation has been tested, and a contrast en-hancement of 2∼3 orders at ps time has been confirmed,as shown in Fig. 2. With an off-axis parabolic (OAP)mirror of 500mm focal length (II-VI Infrared), the laserbeam has been focused to near-diffraction-limited size of16.6µm (horizontal) and 12.3µm (vertical) FWHM, with50% energy enclosed in the Gaussian-fitted focal spot.The measurement is made by directly imaging the focalspot to a 10-bit CCD through a well corrected micro-

TABLE I. The parameters of TW laser system.

Laser parameters ValueCentral wavelength 800nm

Pulse duration 60∼70fsReptition rate 10Hz

Energy on target 500mJEnergy stability 1.5%

Contrast ratio8.1× 106@10ns1.9× 105@10ps2.5× 104@1ps

Focus spot (FWHM)116.6µm (Horizontal)12.3µm (Vertical)

Pointing stability (rms)4.5µrad (Horizontal)5.3µrad (Vertical)

1 Measured with an OAP mirror of 500mm focal length.

−10 −5 0 5 10

1E−7

1E−6

1E−5

1E−4

1E−3

0.01

0.1

1

time(ps)In

tens

ity(a

. u.)

After XPWWithout XPW

FIG. 2. Contrast improvement by using XPW technology be-fore other parameters are optimized. The ps contrasts aremeasured by a third-order autocorrelator with XPW gener-ation (red line) and without XPW generation (blue dashedline).

x (µm)

y (µ

m)

0 10 20 30 40 50

50

40

30

20

10

0

FIG. 3. The horizontal and vertical focal sizes are measuredto be 16.6µm and 12.3µm (FWHM) with f=500mm OAPmirror, respectively, and the 50% energy is enclosed in theGaussian-fitted focal spot. The RMS fluctuation of the focalspot is simultaneously measured to be 2.26µm (horizontal)and 2.65µm (vertical) respectively.

scope objective lens. The RMS fluctuation of the focalspot has been measured to be 2.26µm (horizontal) and2.65µm (vertical) respectively, corresponding to 4∼5µrad

3

angular pointing fluctuation.Obtainable pulse duration of our laser system is

about 60∼70fs, due to the residual high-order dispersionswithin the laser system. An upgradation with a new 25fsfront-end is planned later this year, and this may lead tomuch shorter pulse duration.

B. Linear accelerator

The 45MeV linac, as shown in Fig. 4, has two maincomponents: a modified version of the BNL/KEK/SHItype 1.6 cell photocathode radio-frequency gun and a 3mSLAC-type traveling wave accelerating section. A UVlaser pulse from the TW laser system irradiates the cath-ode nearly perpendicularly to generate an electron pulse.Solenoids are used for emittance compensation and beamenvelope control. And the manipulation of beam canalso be achieved by a pair of quadrupole triplets in-stalled after the accelerating section. Utilizing the 20MWRF power, the maximum electron energy can reach to45MeV with energy fluctuation of 1% and energy spreadof 1% [2]. Several beam diagnostics have been insertedinto the beam line, including beam position monitors,high-resolution YAG screens, charge detectors (Faradaycups and ICT), an S-band RF deflecting cavity [16] andan energy spectrometer.

RF photocathode gun

SolenoidsAccel. sectionQuadrupole triplets

LPA chamber

FIG. 4. The 45MeV linac, which including a photocathodegun, solenoids, a 3m accelerating section, quadrupole tripletsand a deflecting cavity.

C. Experimental system

The experimental system have several main compo-nents, including a vacuum interaction chamber, radia-tion shielding, control and data acquisition (DAQ) andplasma sources. The vacuum interaction chamber hasbeen properly designed to minimize the deformation dur-ing the vacuum pumping down process. Inside the cham-ber, all optical components sit on a reinforced opticalbreadboard which is locked down to an optical table un-der the chamber through bellows. Figure 5 shows themain components of the experiment station. Variousgas jet nozzles of Laval type have been designed andmanufactured. Gas density profiles are measured by in-terferometry for Argon gas. Figure 6 shows an off-line

Compressor

LPA chamber

Sheilding

FIG. 5. The experiment station and the compressor in theexperimental hall.

density measurement platform and two different gas noz-zles. A compact permanent magnetic spectrometer has

(b)

(c)(a)

FIG. 6. The development of plasma source. (a) The off-linegas density measurement platform; (b) A conical gas nozzle;(c) A 1mm×10mm slit gas nozzle.

been carefully designed to measure electron beams with abroad energy range and large momentum spread in typ-ical laser plasma experiment. The magnet has a 15mmpole gap and has the magnetic field near 1T. Figure 7shows the simulated and measuremental data along theelectron injection axis.

−80 −40 0 40 800

0.2

0.4

0.6

0.8

1

The injected direction(mm)

Th

e m

ag

ne

tic

fie

ld(T

) Simulation

Measurement

( ) ( )a b

FIG. 7. The compact permanent magnetic spectrometer. (a)The designed and measured magnetic field along the electroninjection axis; (b) The magnetic energy spectrometer.

III. EXPERIMENTAL RESULTS

Next we present our recent experimental results onlaser wakefield acceleration. Figure 8 shows a schematicof our experimental arrangement (a) and the actual setup

4

( )b

ICTScreen

SlitDe ect Magnetfl

High-Energy detection

Low-Energy detection

5TW Laser800nm@60fs

Interferometer

10×Microscope

Mirror

Relay Image

Imaging Spectrometer

Shadowgraph

Main beam

Probe beam

Nozzle

Wedge

( )a

FIG. 8. A schematic of our experimental arrangement (a) and the experimental setup inside the chamber (b). The main laserpulse is sent into the interaction chamber after compression, and then is focused close to the edge of gas jet nozzle by an OAPmirror. A probe beam, split off from the main beam, is used for interferometry and shadowgraph. The electron beam spectrum,the beam profile and the beam charge can be achieved by a compact magnetic spectrometer and the scintillating screens.

inside the chamber (b). After compression, the mainlaser pulse is sent into the interaction chamber, and thenis focused close to the edge of a 2mm-diameter heliumgas jet nozzle by an OAP mirror with a focal length of500mm. A Mach-Zehnder interferometer combined witha delay line is set up to measure the plasma electrondensity profile with a probe beam split off from the mainbeam. Figure 9 shows a typical interferogram at 3.3nsdelay (a) and the corresponding plasma density obtainedby Abel inversion (b).A scintillating screen (Mitsubishi: PI-200) is inserted

into the beam path to diagnose the beam profile and thebeam charge simultaneously. A 38µm-thick aluminumfoil is placed before the screen to prevent the transmittedlaser light. A magnetic spectrometer can be moved inand out the beam path by a translation stage to measurethe energy spectrum. The spectrometer is equipped withtwo scintillating screens on two sides to monitor the lowenergy and the high energy range. By running the laser

( )b( )a

FIG. 9. (a) The interferometer fringes at 3.3ns delay witha helium back pressure of 0.95Mpa; (b) The correspondingplasma density averaged over the longitudinal direction ob-tained by Abel inversion, which the shaded regions correspondto the standard deviation of the average over the interfero-gram in the longitudinal direction.

at 5TW 60fs, we have obtained 10∼40MeV high qualitymonoenergetic electron beams at a plasma density near

ne = 5 × 1019cm−3. Details of the measurements of theenergy spectrum, the electron beam profile and the beamcharges are presented as follows.

A. Electron energy spectrum and beam profile

In the energy spectrum measurement, a 3mm-widerectangular tungsten slit is introduced into the beampath right before the spectrometer, which gives an un-certainty on the measured energy less than 0.2MeV for a25MeV electron beam. The typical electron beam energyobtained in our experiment is in the range of 10∼40MeV.Here we show two examples. In Fig. 10(a), the electronbeam has a peak energy of 25.7MeV and a FWHM rel-ative energy spread of 4.8%. The vertical divergence ofthe beam is 2.3mrad FWHM. In Fig. 10(b), the electronbeam has a peak energy of 39.2MeV and a FWHM rel-ative energy spread of 8%. The vertical divergence ofthe beam is 10.3mrad FWHM. The cut-off energy of thisbeam is 49MeV, as marked by the green dashed line.A typical electron beam profile is shown in Fig. 11(a).This profile is obtained by a PI-200 screen placed 431mmdownstream the gas jet, with a tilted angle of (61±2)◦

from the beam axis. We also put a few metal wireswith different diameters right before the screen to acquirehigh-resolution electron radiography. Occasionally, mul-tiple electron beams can be generated simultaneously in asingle shot, and an example with three different beamletsis shown in Fig. 11(b).

B. Charge calibration and beam charge

measurement

To measure the electron beam charge accurately, weadopted a widely accepted method based on sensitivescintillating screens. Several authors provide calibrationdata on different scintillating screens [17, 18] using elec-

5

27 1. 20 6. 15 3. 13 4. 8 7. 7 1.34 4.

- .7 80

7 8.15 6.

Energy(MeV)Div

erg

ence(m

rad)

E =25.7MeV, E /E =4.8%peak FWHM peakΔ

34.5 39.7 48.4 66 103

−13

−6.5

0

6.5E =39.2MeV, E /E =8.0%peak FWHM peakΔ

Div

erg

ence(m

rad)

Energy(MeV)

4

(a)

(b)

FIG. 10. Two examples of monoenergetic electron beams:(a) The peak energy of 25.7MeV with a FWHM relative en-ergy spread of 4.8%. (b) The peak energy of 39.2MeV witha FWHM relative energy spread of 8%. The cut-off energymarked by the green dashed line is 49MeV, while the laserinjection position is marked by the red dotted line.

x( m)m

y(

m)

m

0 10 20 30 40 50

0

10

20

30

40

50

Shadow of rectangular slit

Shadow of wires(160 m/100 m)μ μ

x(mm)

y(m

m)

0 11 22 33 44 55

0

11

22

33

44

55

( )a ( )b

FIG. 11. Electron beam profiles on the scintillating screens:(a) The FWHM beam size with 5.53mm (horizontal) and4.12mm (vertical), contributing to the average divergence of11mrad. The images of a rectangular slit and two metal wires(100µm and 160µm in diameter) are simultaneously producedby this accelerated beam. (b) Three different beamlets in asingle shot.

tron beams from linac. They found two types of screens(PI-200 and Drz-high) are more sensitive than the Lanexregular screen. However directly adopting their calibra-tion data in our experiment may cause some ambiguitiesbecause the exact setups and the properties of the CCDcameras are different from experiment to experiment. Toavoid these ambiguities, we performed a similar calibra-tion procedure using our in-house linac system. A rela-tive shorter exposure time (100µs) is used in our calibra-tion to reduce the background noise.The electron beam energy of the linac is fixed at

30MeV with a bunch duration of about 1ps, and anICT (Bergoz ICT-055-070-05:1) is used to benchmarkthe electron charge. The electron charge is kept belowthe saturation threshold of the scintillating screens[17].A cooled 16-bit CCD camera (Apogee: Alta U2) witha Nikon lens (focal length 60mm and F number 1/2.8)is used to record the signal, and a band pass filter at546 nm is also applied to reduce the background light.In Fig. 12, we plot the calibration data for three dif-ferent screens. We found that the PI-200 and Drz-high

screens are about three times more sensitive than theLanex regular screen. The measured sensitivity of PI-200 is 1.67×109 counts/sr/pC, while the sensitivity ofDrz-high is 1.42×109 counts/sr/pC. Based on our calibra-tion, the charge for the electron beam shown in Fig. 11(a)is 10.6pC, and the charges for energy measurements arearound 0.2pC (shown in Fig.10(a) and (b)).

0 20 40 60 80 100 1200

1

2

3

4

5

6x 10

7

Charge (pC)

Cou

nts

Drz−highLanex regularPI−200

1.67×109

1.42×109

3.8×108

FIG. 12. Absolute charge calibration of three different screens(red for Drz-high, blue for Lanex regular and black for PI-200) using electron beams from the linac. The makers showthe measuremental data, and the gradients of the best fit linesare also provided.

IV. CONCLUSIONS

A unique facility for laser plasma physics and advancedaccelerator research has been built recently at TsinghuaUniverstiy. This system is based on Tsinghua Thomsonscattering X-ray source (TTX), which combining an ul-trafast TW laser with a synchronized 45MeV high bright-ness linac. In our recent laser wakefield acceleration ex-periments, we have obtained 10∼40MeV high quality mo-noenergetic electron beams by running the laser at 5TWpeak power. Under certain conditions, very low relativeenergy spreads of a few percent can be achieved, whichcloses to the best published results [12–14]. Absolutecharge calibration for three different scintillating screenshas also been performed using our linac system, and thetypical charge of accelerated electron beams is in the pic-ocoulomb range.

ACKNOWLEDGMENTS

The authors would like to thank Dr. Z. Y. Wei, Dr. H.Teng, Dr. Z. H. Wang and Dr. J. L. Ma from Institute ofPhysics (CAS) for their great support on the laser systemconstruction, Dr. J. Wang, Dr. H. -H Chu and Dr. S. -Y.Chen from National Central Universtiy (Taiwan) for theirvaluable information on the laser system optimization.

6

We also thank to Dr. J. Zhang from China Institute ofAtomic Energy for the use of their contrast measurement

equipment.

[1] Chuangxiang Tang, Wenhui Huang, Renkai Li, et al.,Nucl.Instr.and Meth.A, 2009, 608: S70

[2] Yingchao Du, Lixin Yan, Jianfei Hua, et al., Rev. Sci.Instrum., 2013, 84: 053301

[3] T. Tajima and J. M. Dawson, Phys. Rev. Lett., 1979, 43:267

[4] Pisin Chen, J. M. Dawson, Robert W. Huff, et al., Phys.Rev. Lett., 1985, 54: 693

[5] F. S. Tsung, Ritesh Narang, W. B. Mori, et al., Phys.Rev. Lett., 2004, 93: 185002

[6] W. P. Leemans, B. Nagler, A. J. Gonsalves, et al., Nat.Phys., 2006, 2: 696

[7] Nasr A. M. Hafz, Tae Moon Jeong, IL Woo ChoI, et al.,Nat. Photonics, 2008, 2: 571

[8] C. E. Clayton, J. E. Ralph, F. Albert, et al., Phys. Rev.Lett., 2010, 105: 105003

[9] J. S. Liu, C. Q. Xia, W. T. Wang, et al., Phys. Rev. Lett.,2011, 107: 035001

[10] Hyung Taek Kim, Ki Hong Pae, Hyuk Jin Cha, et al.,Phys. Rev. Lett., 2013, 111: 165002

[11] Xiaoming Wang, Rafal Zgadzaj, Neil Fazel, et al., Nat.Commun., 2013, 4: 1988

[12] S. M. Wiggins, R. C. Issac, G. H. Welsh, et al., PlasmaPhys. Control. Fusion, 2010, 52: 124032

[13] B. B. Pollock, C. E. Clayton, J. E. Ralph, et al., Phys.Rev. Lett., 2011, 107: 045001

[14] A. Buck, J. Wenz, J. Xu, et al., Phys. Rev. Lett., 2013,110: 185006

[15] Yan Li-xin, Du Ying-chao, Du Qiang, et al., Chin. Phys.C, 2009, 33: 154

[16] Jiaru Shi, Huaibi Chen, Chuangxiang Tang, et al., Pro-ceedings of PAC09, 2009, 3429

[17] A. Buch, K. Zeil, A. Popp, et al., Rev. Sci. Instrum.,2010, 81: 033301

[18] Y. C. Wu, B. Zhu, K. G. Dong, et al., Rev. Sci. Instrum.,2012, 83: 026101